Pith. sign in

REVIEW 3 major objections 4 minor 66 references

Optimal Nonergodic Primal-Dual Complexity of Efficient Inexact Parameter-Free Augmented Lagrangian Methods

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read The paper proves that standard inexact augmented Lagrangian methods can compute verifiable approximate KKT points for linearly constrained convex composite problems at the optimal O(1/epsilon) first-order complexity, including…

desk verdict The O(1/ε) claim is likely right but the printed algebra doesn't support it: the max in the PF-AR bound belongs inside the square root, so Theorem 2.8(b) as written yields O(ε^{-3/2}), not O(1/ε). read the letter →

arxiv 2608.03170 v2 pith:3O3Z7NZF submitted 2026-08-04 math.OC

Arnesh Sujanani , Saeed Ghadimi , Henry Wo{l}kowicz This is my paper · ORCID
classification math.OC MSC 65K0549M3790C0665K1090C30
keywords augmentedLagrangianfirst-ordermethodsparameter-freeoptimalcomplexityverifiableterminationprimal-dualsolutionKKTpointlinearlyconstrainedconvexoptimization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether the classical inexact augmented Lagrangian (AL) architecture can find verifiable approximate KKT points for linearly constrained convex composite problems at the optimal $\mathcal O(1/\epsilon)$ first-order complexity, without proximal regularization, ergodic averaging, or prior knowledge of problem constants. It answers yes, constructing three methods, O-IAL, OPF-IAL, and APF-IAL, that attain $\mathcal O(1/\epsilon)$ gradient-evaluation complexity for finding an $(\epsilon,\epsilon)$-approximate primal-dual solution. Two of the methods are parameter-free, and all three give nonergodic guarantees, with a last-iterate guarantee for APF-IAL. In the strongly convex case the same methods reach $\mathcal O(\epsilon^{-1/2}\log(\epsilon^{-1}))$ without knowing the strong convexity modulus. The paper also reports speedups frequently ranging from 5 to 50 times over a representative proximal AL method on six problem classes.

What carries the argument

The central object is PF-AR, a parameter-free accumulative regularization method for convex composite minimization that produces verifiable stationarity certificates, i.e., a pair $(\hat x,\hat v)$ with $\hat v\in\nabla\psi_s(\hat x)+\partial\psi_n(\hat x)$ and $\|\hat v\|\le\epsilon$ in $\mathcal O(1/\epsilon)$ gradient evaluations. PF-AR repeatedly calls a subroutine AR-L with doubling diameter guesses; AR-L uses a backtracking line-search and a termination criterion based on a true subgradient norm rather than a gradient mapping, which is essential for the AL analysis. The AL outer loop then combines approximate dual stationarity with the full multiplier update to obtain the telescoping estimate that bounds feasibility residuals by $\|p^*-p_0\|^2$. For strongly convex problems, a restarted FISTA scheme, R-FISTA, supplies the inner solver to reach $\mathcal O(\epsilon^{-1/2}\log(\epsilon^{-1}))$ complexity.

What would settle it

Run OPF-IAL on a family of linearly constrained convex instances with known smoothness, domain diameter, and penalty parameter set to $c=1/\epsilon$, measuring the total number of gradient evaluations needed to reach $\|Az-b\|\le\epsilon$ and $\|r\|\le\epsilon$ for $\epsilon=10^{-3},10^{-4},10^{-5},10^{-6}$. If the empirical growth is worse than $\mathcal O(1/\epsilon)$ on a log-log plot, or if the output triple ever fails the containment and residual conditions of (2.2), then the central claim would be disproved.

Watch

Extended reading notes

Core claim

The central claim is that the standard, unregularized inexact AL framework is sufficient for optimal verifiable primal-dual complexity. For problems of the form $\min \{\psi_s(z)+\psi_n(z) : Az=b\}$ with convex smooth $\psi_s$ and closed proper convex $\psi_n$ on a compact domain, the paper constructs three AL methods that output a triple $(z,p,r)$ with $r\in \nabla\psi_s(z)+\partial\psi_n(z)+A^*p$, $\|r\|\le\epsilon$, and $\|Az-b\|\le\epsilon$, in $\mathcal O(1/\epsilon)$ gradient evaluations. The key is to maintain approximate dual stationarity at every outer iteration; then the multiplier update $p_{k+1}=p_k+c(Az_{k+1}-b)$ yields a telescoping dual-distance estimate that controls primal feasibility without ergodic averaging. This improves prior verifiable AL bounds of $\mathcal O(\epsilon^{-4/3})$, $\mathcal O(\epsilon^{-3/2})$, $\mathcal O(\epsilon^{-7/4})$, and $\mathcal O(\epsilon^{-2})$, and removes the logarithmic factor from proximal AL guarantees.

Load-bearing premise

The entire optimal-complexity result rests on the assumption that every augmented-Lagrangian subproblem can be solved by an accelerated subroutine whose error after $k$ steps shrinks like a constant over $k^2$, using a smoothness estimate that is only a constant factor larger than the true one.

Editorial extensions

If this is right

  • For linearly constrained convex composite problems, optimal $\mathcal O(1/\epsilon)$ first-order complexity can be achieved with standard, unregularized AL subproblems and a directly verifiable stopping rule.
  • Two of the proposed methods, OPF-IAL and APF-IAL, remove the need to know the domain diameter or the distance to an optimal multiplier, so the optimal rate is attainable without problem-dependent tuning.
  • In the strongly convex setting all three methods attain $\mathcal O(\epsilon^{-1/2}\log(\epsilon^{-1}))$ without knowledge of the strong convexity parameter.
  • The APF-IAL variant adaptively increases the penalty parameter and subproblem tolerances, and it provides a last-iterate guarantee, not only a best-iterate guarantee.
  • On six problem classes, including elastic-net regression, group-sparse SVMs, and a quantum semidefinite program, the methods were often 5 to 50 times faster than the ProxALM baseline under identical termination criteria.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the outer-loop telescoping argument depends only on small dual-stationarity residuals, so any inner solver providing such certificates at optimal cost could carry the same $\mathcal O(1/\epsilon)$ AL complexity, potentially extending the result to other composite structures beyond compact domains.
  • Editorial inference: the PF-AR stationarity-certificate subroutine is likely reusable as a standalone parameter-free solver for composite minimization with verifiable stopping, since its guarantees do not rely on the AL outer loop.
  • Editorial inference: a natural stress test is whether the optimal bound persists when the smoothness of the AL subproblem grows with the penalty parameter $c$; the analysis suggests it does as long as an accelerated method satisfying the constant-factor Lipschitz estimate is available, but the paper does not test this regime numerically.
  • Editorial inference: the same design may transfer to stochastic or distributed oracles whenever the inner solver can maintain approximate dual stationarity, since the outer complexity proof is agnostic to how the inner certificate is produced.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes three inexact augmented Lagrangian methods (O-IAL, OPF-IAL, APF-IAL) for linearly constrained convex composite problems with a compact nonsmooth domain. The central claim is that these methods attain an optimal O(1/epsilon) first-order complexity for finding verifiable (epsilon, epsilon)-approximate primal-dual (KKT) solutions in the convex case, with two parameter-free variants, and a near-optimal O(epsilon^{-1/2} log(1/epsilon)) bound in the strongly convex case. The algorithms preserve the standard unregularized AL subproblem structure and use two new parameter-free inner solvers: PF-AR for convex subproblems and R-FISTA for strongly convex subproblems. Numerical experiments on six problem classes compare O-IAL and APF-IAL against the proximal AL method of Lu and Mei, reporting frequent 5-50x speedups. The proofs combine a telescoping dual-distance argument for outer iterations with a parameter-free accelerated regularization analysis for inner iterations.

Significance. If the claimed bounds are correct, the paper settles an interesting structural question: standard, unregularized inexact AL methods, without proximal regularization, ergodic averaging, or knowledge of problem constants, can in principle match the optimal verifiable KKT complexity previously achieved only by specialized or proximal variants. The paper also contains a genuinely useful parameter-free composite stationarity solver, PF-AR, and a simplified restarted strongly convex solver, R-FISTA. The numerical study is broad and includes an SDP application, and the code is publicly linked. However, the advertised O(1/epsilon) conclusion rests on a load-bearing algebraic bound that is misstated in the printed theorems, and one key contraction estimate in the AR-L proof is not justified by the stated assumption; both issues are repairable but currently prevent the central claim from following as written.

major comments (3)
  1. [Theorem 2.8(b), Proposition 2.9(b), and Theorem 2.13 / Proposition 2.17(b)] The displayed inner complexity bound (2.27) is not implied by Theorem 2.6. Since PF-AR has complexity O(sqrt(L_sub D / epsilon_inner)) with L_sub = Lbar + c||A||^2 and epsilon_inner = min{c eps^2/(4D), rho}, the correct inner bound is O(sqrt((Lbar + c||A||^2) * max{D^2/(c eps^2), D/rho})), with the max inside the square root. As printed, the max is outside, giving O(sqrt(Lbar + c||A||^2) * max{D^2/(c eps^2), D/rho}). Substituting c = O(1/eps) into the printed bound yields O(eps^{-3/2}) for eps = rho, not O(1/eps). The same misplaced max appears in (2.24), (2.25), and in the OPF-IAL bounds (2.47) and Theorem 2.13. The advertised optimal O(1/eps) complexity therefore does not follow from the manuscript as written, although it does follow after moving the max inside the square root and then substituting c = 1/eps, rho = eps.
  2. [Assumption 2.3 and Appendix A, Eqs. (A.17)-(A.19)] The contraction estimate used to prove Theorem 2.4 is not justified. In (A.17)-(A.19) the authors conclude that (1/N_s) sqrt(2 c_A Lbar / sigma_s) <= 1/8 using the bound (A.18). But (A.18) is an upper bound on N_s, namely N_s <= 8 sqrt(2 c_A Lbar / sigma_s), so the displayed inequality has the wrong direction. To obtain a 1/8 contraction, one needs a lower bound on the number of inner iterations, e.g., a lower bound on the Lipschitz estimate L_s^k. The stated Assumption 2.3 only provides L_s^k <= c_A Lbar. This gap affects the proof of AR-L and hence the PF-AR complexity in Theorem 2.6, which is the engine behind all three AL methods. The issue is likely repairable by strengthening Assumption 2.3 to also require L_s^k >= Lbar / c_A, which standard backtracking or universal-gradient implementations do satisfy, but it must be stated and used consistently.
  3. [Proposition 2.12] The outer-iteration bound (2.36) is stated using Pi(p0), but the proof says 'let p* be an arbitrary optimal solution' and then uses ||p* - p0||^2. To obtain Pi(p0) one must choose p* to minimize the distance to p0, or take an infimum over optimal multipliers. As written the transition from 'arbitrary p*' to Pi(p0) is not immediate. This is a local fix, but it should be clarified.
minor comments (4)
  1. [Section 2.4, Theorem 2.20] The APF-IAL bound (2.50) does not contain the misplaced-max error found in O-IAL and OPF-IAL: the inner PF-AR estimate in Proposition 2.22(b) correctly places the tolerance inside the square root. It would be helpful to state explicitly that the O(1/eps) claim for APF-IAL is not affected by the correction needed in (2.27).
  2. [Section 4] The numerical experiments set c = 10 (or c1 = 10) while the theoretical optimal complexity requires c = O(1/eps) with eps = 1e-5. The paper acknowledges testing outside the compact-domain assumption in some experiments but does not mention this deviation from the parameter regime of Theorem 2.8. A brief remark would help readers interpret the experiments.
  3. [Appendix B, Lemma B.1 and Proposition 3.2] The strongly convex analysis relies on Lemma B.1, whose proof is omitted with a reference to prior work [43, 44]. Given that Lemma B.1 is load-bearing for the near-optimal strongly convex claims, the paper should either reproduce the proof or state very precisely which statements in [44] are being imported.
  4. [Assumption 2.3] The terminology 'estimate of the Lipschitz constant' in Assumption 2.3 should specify whether L_s^k is an upper estimate, a lower estimate, or both up to a universal factor. The current wording 'L_s^k <= c_A Lbar' is compatible with the proof only if a lower bound of similar type is also assumed.

Circularity Check

3 steps flagged · score 4.0 of 10

No fitted-input circularity, but APF-IAL and R-FISTA import load-bearing lemmas wholesale from the authors' prior work [43,44]; the printed O(1/epsilon) bound also contains an internal max/sqrt algebra slip.

  1. self citation load bearing [Lemma 2.21, Section 2.4.1 (Proof of Theorem 2.20)]
    "The following key result from [43] is presented with minor changes in notation. It will be instrumental in showing that the sequence of Lagrange multipliers {p k} is bounded."

    Lemma 2.21 is the sole stated source of the uniform bound on {p_k} in Proposition 2.24. That bound produces the feasibility estimate (2.58), the penalty-parameter bound (2.59), and the outer-iteration count (2.64) used in Theorem 2.20, so the APF-IAL last-iterate O(1/epsilon) claim depends on a result [43] by the same first author that is stated without proof here. The cited lemma is not machine-checked in this manuscript, so the load-bearing step is not verified by the present derivation.

  2. self citation load bearing [Appendix B.1, Lemma B.1]
    "The following lemma presents key properties of the iterates generated during the ℓ-th cycle of R-FISTA. Its proof is not given as it closely resembles the proofs of Lemmas A.3 and A.4 in [43]."

    Lemma B.1 supplies the key inequalities (B.2)-(B.4), including the lower bound A_j L_j >= max{j^2/4, (1+Q_ell^{-1})^{2(j-1)}}, from which Proposition 3.2(a) derives the cycle iteration count. Since the proof of Lemma B.1 is explicitly omitted and attributed to [43], the R-FISTA complexity underlying Theorem 3.3, and hence the strongly convex bounds of Theorems 3.5 and 3.7, rests on an unverified-in-this-paper self-citation rather than on a contained derivation.

1 more flagged steps
  1. self citation load bearing [Proof of Proposition 3.2(c), Appendix B.2]
    "It can easily be seen that Lemma B.2 and the fact that ˜µ=µ ℓ−1 is in the interval (0,¯µ] imply that Lemmas A.2-A.6 in [44] hold. The next lemma thus restates Lemma A.6 of [44]..."

    Proposition 3.2(c) is the no-restart result needed to bound the number of R-FISTA cycles in Proposition C.1(a). Its proof says that Lemmas A.2-A.6 of [44] hold and then restates Lemma A.6 of [44] as Lemma B.3; reference [44] is by the same authors and is not proved in this manuscript. This makes the near-optimal strongly convex complexity claim load-bearing on a chain of self-citations.

full rationale

The paper's central convex O(1/epsilon) analysis is not circular: O-IAL and OPF-IAL derive outer-iteration counts from the telescoping inequality (2.33), and the inner PF-AR complexity is quoted from Theorem 2.6, which explicitly depends on Assumption 2.3 and on the external method of Lan et al. [16]. No parameter is fitted to the target output, and termination criterion (2.2) is used as the definition of the output rather than as an imposed conclusion. However, the manuscript is not fully self-contained. The bounded-multiplier lemma used by APF-IAL (Lemma 2.21) is imported from [43] with the same first author, and the R-FISTA cycle analysis rests on Lemma B.1, whose proof is omitted and assigned to [43], plus Lemmas A.2-A.6 of [44]; these are load-bearing for the last-iterate and strongly convex complexity claims. Separately, but not a circularity, the displayed complexity in (2.24)-(2.27) puts 'max' outside the square root, so the printed algebra would give O(epsilon^{-3/2}) rather than the advertised O(1/epsilon); correcting the 'max' to the inside of the square root, as Theorem 2.6 requires, would restore the stated bound. That is a correctness issue, not an equivalence-by-construction. The overall score reflects the unproved imported self-cited lemmas, not a reduction of the derivation to its inputs.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new physical or mathematical entities. Its central claims rest on standard convex-analysis assumptions plus an algorithmic assumption (Assumption 2.3) about available accelerated subsolvers. The main non-self-contained parts are the deferred proofs imported from the authors' own prior papers [43,44], and the hand-chosen penalty and adaptivity parameters that affect the complexity constants.

free parameters (2)
  • Penalty parameter c (O-IAL and OPF-IAL) = O(1/epsilon) in Theorem 2.8(b); c = 10 in experiments
    Hand-chosen algorithmic parameter, not fitted to data. The O(1/epsilon) claim is sensitive to c; as printed, the proof uses c = O(1/epsilon) but the per-iteration bound appears to be missing a square root.
  • APF-IAL parameters c1, alpha, omega, eps~1 = c1 = 10, alpha = 1.1, omega = 1.5, eps~1 = 10 in experiments
    User-chosen hyperparameters of the adaptive scheme. The asymptotic bound treats them as constants; they are not fitted to data but do affect constants in the bound.
assumptions (6)
  • domain assumption Assumption 2.1: existence of optimal primal-dual pair (z*, p*) satisfying the KKT conditions
    Used throughout for the KKT criterion and the dual-distance telescoping. Compact feasible domain gives primal solvability, but an optimal multiplier is assumed to exist.
  • domain assumption psi_n is closed proper convex with compact domain, diameter D finite
    Drives the bound ||z - z*|| <= D, the PF-AR cost, and several Markov/Cauchy-Schwarz arguments. All three AL methods rely on it.
  • domain assumption psi_s is convex and L-smooth
    Ensures the AL subproblem is (L + c||A||^2)-smooth and that the backtracking line search terminates.
  • ad hoc to paper Assumption 2.3: existence of an accelerated solver A with function-gap guarantee (2.8) and bounded Lipschitz estimate
    PF-AR's optimal O(sqrt(LD/epsilon)) complexity is conditional on this algorithmic assumption. R-FISTA is claimed to satisfy it, but its key lemmas are deferred to [44].
  • domain assumption Assumptions 2.18 and 2.19 for APF-IAL: psi_n is Lipschitz on dom psi_n and there exists zbar in int(dom psi_n) with Azbar = b
    Used through Lemma 2.21 (imported from [43]) to bound the multiplier sequence and enable the last-iterate guarantee for APF-IAL.
  • standard math External results from [16], [43], and [44] are correct
    AR-L lemmas and R-FISTA lemmas are cited rather than reproved; the strong-convexity and bounded-multiplier analyses depend on these prior works.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Optimal Nonergodic Primal-Dual Complexity of Efficient Inexact Parameter-Free Augmented Lagrangian Methods." pith.science (2026). https://pith.science/paper/3O3Z7NZF

@misc{pith2026260803170,
  author       = {Pith},
  title        = {Pith review of: Optimal Nonergodic Primal-Dual Complexity of Efficient Inexact Parameter-Free Augmented Lagrangian Methods},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3O3Z7NZF}},
  note         = {Machine review of arXiv:2608.03170}
}
abstract

Augmented Lagrangian (AL) methods are a classical framework for constrained optimization, but for directly verifiable approximate KKT points, known first-order complexity bounds for standard inexact AL methods are suboptimal, while the best known proximal augmented Lagrangian (PAL) bounds retain an additional logarithmic factor. We consider linearly constrained convex composite problems with a smooth convex term and a possibly nonsmooth closed proper convex term with compact domain. We develop three inexact AL schemes that preserve the standard AL subproblem structure and attain the optimal primal-dual complexity $\mathcal O(\epsilon^{-1})$ in the convex setting, improving prior AL bounds of $\mathcal O(\epsilon^{-4/3})$, $\mathcal O(\epsilon^{-7/4})$, and $\mathcal O(\epsilon^{-2})$, and removing the logarithmic factor from PAL guarantees. Two variants are parameter-free, and all three admit nonergodic guarantees, including a stronger last-iterate guarantee for one variant. These results show that proximal regularization, ergodic averaging, and prior knowledge of problem-dependent constants are not intrinsic requirements for attaining optimal verifiable primal-dual complexity within the standard AL framework. A key ingredient is a parameter-free accelerated method that computes verifiable stationarity certificates for the standard, unregularized AL subproblems with optimal complexity. In the strongly convex setting, our methods attain near-optimal complexity $\mathcal O(\epsilon^{-1/2}\log(\epsilon^{-1}))$, with two parameter-free variants. Numerical experiments on six problem classes, including elastic-net least-squares regression, group-sparse Huberized support vector machines, and a quantum semidefinite program (SDP), demonstrate substantial computational advantages over a representative PAL method, with speedups frequently ranging from $5$ to $50$ times.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 52 canonical work pages

  1. [44]

    Journal of Optimization Theory and Applications , volume=

    FISTA restart using an automatic estimation of the growth parameter , author=. Journal of Optimization Theory and Applications , volume=. 2025 , publisher=

    work page 2025

  2. [43]

    Journal of Optimization Theory and Applications , volume=

    Efficient parameter-free restarted accelerated gradient methods for convex and strongly convex optimization , author=. Journal of Optimization Theory and Applications , volume=. 2025 , publisher=

    work page 2025

  3. [1]

    2013 , publisher=

    Perturbation analysis of optimization problems , author=. 2013 , publisher=

    2013

  4. [2]

    Censor and and W

    Y. Censor and and W. Moursi and T. Weames and H. Wolkowicz. Regularized Nonsmooth N ewton Algorithms for Best Approximation with Applications. 2022 submitted

    work page 2022

  5. [3]

    2024 , eprint=

    Optimal and parameter-free gradient minimization methods for convex and nonconvex optimization , author=. 2024 , eprint=

    work page 2024

  6. [4]

    SIAM Journal on Optimization , volume=

    Accelerated first-order methods for convex optimization with locally Lipschitz continuous gradient , author=. SIAM Journal on Optimization , volume=. 2023 , publisher=

    work page 2023

  7. [5]

    SIAM journal on optimization , volume=

    Iteration-complexity of first-order augmented Lagrangian methods for convex conic programming , author=. SIAM journal on optimization , volume=. 2023 , publisher=

    work page 2023

  8. [6]

    Mathematical Programming , volume=

    Iteration complexity of inexact augmented Lagrangian methods for constrained convex programming , author=. Mathematical Programming , volume=. 2021 , publisher=

    work page 2021

Show all 66 references
  1. [7]

    International Conference on Artificial Intelligence and Statistics , pages=

    Rate-improved inexact augmented Lagrangian method for constrained nonconvex optimization , author=. International Conference on Artificial Intelligence and Statistics , pages=. 2021 , organization=

    2021

  2. [8]

    Computational Optimization and Applications , volume=

    Convergence rate of inexact augmented Lagrangian method with practical relative error criterion for composite convex programming , author=. Computational Optimization and Applications , volume=. 2025 , publisher=

    2025

  3. [9]

    arXiv preprint arXiv:2602.17878 , year=

    Improved analysis of restarted accelerated gradient and augmented Lagrangian methods via inexact proximal point frameworks , author=. arXiv preprint arXiv:2602.17878 , year=

  4. [10]

    SIAM Journal on Optimization , volume =

    Lu, Zhaosong and Mei, Sanyou , title =. SIAM Journal on Optimization , volume =. 2023 , doi =. https://doi.org/10.1137/22M1500496 , abstract =

    2023 doi

  5. [11]

    Mathematical Programming Computation , volume=

    An inexact proximal augmented Lagrangian framework with arbitrary linearly convergent inner solver for composite convex optimization , author=. Mathematical Programming Computation , volume=. 2021 , publisher=

    2021

  6. [12]

    SIAM journal on optimization , volume=

    Reducing the complexity of two classes of optimization problems by inexact accelerated proximal gradient method , author=. SIAM journal on optimization , volume=. 2023 , publisher=

    2023

  7. [13]

    arXiv preprint arXiv:2509.02894 , year=

    A proximal augmented Lagrangian method for nonconvex optimization with equality and inequality constraints , author=. arXiv preprint arXiv:2509.02894 , year=

    arXiv

  8. [14]

    Advances in Neural Information Processing Systems , volume=

    An inexact augmented Lagrangian framework for nonconvex optimization with nonlinear constraints , author=. Advances in Neural Information Processing Systems , volume=

  9. [15]

    INFORMS Journal on Optimization , volume=

    Augmented Lagrangian--based first-order methods for convex-constrained programs with weakly convex objective , author=. INFORMS Journal on Optimization , volume=. 2021 , publisher=

    2021

  10. [16]

    arXiv preprint arXiv:2605.17089 , year=

    A preconditioned augmented Lagrangian method for solving semidefinite programming problems , author=. arXiv preprint arXiv:2605.17089 , year=

    arXiv

  11. [17]

    Mathematical Programming , volume=

    Iteration-complexity of first-order augmented Lagrangian methods for convex programming , author=. Mathematical Programming , volume=. 2016 , publisher=

    2016

  12. [18]

    Mathematics of Operations Research , volume=

    On the nonergodic convergence rate of an inexact augmented Lagrangian framework for composite convex programming , author=. Mathematics of Operations Research , volume=. 2019 , publisher=

    2019

  13. [19]

    INFORMS Journal on Optimization , volume=

    First-order methods for constrained convex programming based on linearized augmented Lagrangian function , author=. INFORMS Journal on Optimization , volume=. 2021 , publisher=

    2021

  14. [20]

    Optimization Methods and Software , volume=

    Complexity of first-order inexact Lagrangian and penalty methods for conic convex programming , author=. Optimization Methods and Software , volume=. 2019 , publisher=

    2019

  15. [21]

    Optimization Letters , volume=

    Adaptive inexact fast augmented Lagrangian methods for constrained convex optimization , author=. Optimization Letters , volume=. 2017 , publisher=

    2017

  16. [22]

    arXiv preprint arXiv:1302.6322 , year=

    An augmented Lagrangian method for conic convex programming , author=. arXiv preprint arXiv:1302.6322 , year=

    arXiv

  17. [23]

    Optimization Methods and Software , volume=

    Iteration complexity analysis of dual first-order methods for conic convex programming , author=. Optimization Methods and Software , volume=. 2016 , publisher=

    2016

  18. [24]

    Mathematics of Operations Research , volume=

    Iteration complexity of a proximal augmented Lagrangian method for solving nonconvex composite optimization problems with nonlinear convex constraints , author=. Mathematics of Operations Research , volume=. 2023 , publisher=

    2023

  19. [25]

    arXiv preprint arXiv:2605.22539 , year=

    A conditional-gradient-based single-loop augmented Lagrangian method for inequality constrained optimization , author=. arXiv preprint arXiv:2605.22539 , year=

    arXiv

  20. [26]

    arXiv preprint arXiv:2505.15775 , year=

    New understandings and computation on augmented lagrangian methods for low-rank semidefinite programming , author=. arXiv preprint arXiv:2505.15775 , year=

    arXiv

  21. [27]

    International Conference on Machine Learning , pages=

    A conditional-gradient-based augmented Lagrangian framework , author=. International Conference on Machine Learning , pages=. 2019 , organization=

    2019

  22. [28]

    Mathematical Programming , volume=

    Iteration-complexity of first-order penalty methods for convex programming , author=. Mathematical Programming , volume=. 2013 , publisher=

    2013

  23. [29]

    Journal of Scientific Computing , volume=

    Solving low-rank semidefinite programs via manifold optimization , author=. Journal of Scientific Computing , volume=. 2025 , publisher=

    2025

  24. [30]

    Operations Research , volume=

    A low-rank augmented Lagrangian method for doubly nonnegative relaxations of mixed-binary quadratic programs , author=. Operations Research , volume=. 2026 , publisher=

    2026

  25. [31]

    Mathematical Programming , pages=

    Optimal and parameter-free gradient minimization methods for convex and nonconvex optimization , author=. Mathematical Programming , pages=. 2026 , publisher=

    2026

  26. [32]

    Journal of optimization theory and applications , volume=

    Optimizing the efficiency of first-order methods for decreasing the gradient of smooth convex functions , author=. Journal of optimization theory and applications , volume=. 2021 , publisher=

    2021

  27. [33]

    How to make the gradients small , author=. Optima. Mathematical Optimization Society Newsletter , number=

  28. [34]

    SIAM Journal on Control and Optimization , volume=

    Computational complexity of inexact gradient augmented Lagrangian methods: application to constrained MPC , author=. SIAM Journal on Control and Optimization , volume=. 2014 , publisher=

    2014

  29. [35]

    Journal of Optimization Theory and Applications , volume=

    Universal subgradient and proximal bundle methods for convex and strongly convex hybrid composite optimization , author=. Journal of Optimization Theory and Applications , volume=. 2026 , publisher=

    2026

  30. [36]

    SIAM Journal on Optimization , volume=

    Complexity of a quadratic penalty accelerated inexact proximal point method for solving linearly constrained nonconvex composite programs , author=. SIAM Journal on Optimization , volume=. 2019 , publisher=

    2019

  31. [37]

    Advances in Neural Information Processing Systems , volume=

    A geometric structure of acceleration and its role in making gradients small fast , author=. Advances in Neural Information Processing Systems , volume=

  32. [38]

    A Parameter-Free Restart Scheme with Only a Parallelizable

    Wu, Yue and Grimmer, Benjamin , journal=. A Parameter-Free Restart Scheme with Only a Parallelizable

  33. [39]

    arXiv preprint arXiv:2505.11670 , year=

    An Adaptive and Parameter-Free Nesterov's Accelerated Gradient Method for Convex Optimization , author=. arXiv preprint arXiv:2505.11670 , year=

    arXiv

  34. [40]

    Mathematics of Operations Research , volume=

    AdaBB: Adaptive Barzilai-Borwein method for convex optimization , author=. Mathematics of Operations Research , volume=. 2026 , publisher=

    2026

  35. [41]

    Mathematical Programming , pages=

    A simple uniformly optimal method without line search for convex optimization , author=. Mathematical Programming , pages=. 2025 , publisher=

    2025

  36. [42]

    Sujanani and R.D.C

    A. Sujanani and R.D.C. Monteiro , title =. J. Scientific Computing , year =

  37. [45]

    Mathematics of Operations Research , volume=

    Augmented Lagrangians and Applications of the Proximal Point Algorithm in Convex Programming , author=. Mathematics of Operations Research , volume=. 1976 , publisher=

    1976

  38. [46]

    , author=

    A Low-rank augmented lagrangian method for large-scale semidefinite programming based on a hybrid convex-nonconvex approach. , author=. Mathematical Programming , pages=. 2026 , publisher=

    2026

  39. [47]

    arXiv preprint arXiv:2505.13719 , year=

    cuhallar: A gpu accelerated low-rank augmented lagrangian method for large-scale semidefinite programming , author=. arXiv preprint arXiv:2505.13719 , year=

  40. [48]

    SIAM Journal on Optimization , volume=

    An average curvature accelerated composite gradient method for nonconvex smooth composite optimization problems , author=. SIAM Journal on Optimization , volume=. 2021 , publisher=

    2021

  41. [49]

    Computational Optimization and Applications , volume=

    Average curvature FISTA for nonconvex smooth composite optimization problems , author=. Computational Optimization and Applications , volume=. 2023 , publisher=

    2023

  42. [50]

    SIAM journal on imaging sciences , volume=

    A fast iterative shrinkage-thresholding algorithm for linear inverse problems , author=. SIAM journal on imaging sciences , volume=. 2009 , publisher=

    2009

  43. [51]

    Mathematical programming , volume=

    Gradient methods for minimizing composite functions , author=. Mathematical programming , volume=. 2013 , publisher=

    2013

  44. [52]

    SIAM Journal on Optimization , volume=

    Another look at the fast iterative shrinkage/thresholding algorithm (FISTA) , author=. SIAM Journal on Optimization , volume=. 2018 , publisher=

    2018

  45. [53]

    Foundations of computational mathematics , volume=

    Fast first-order methods for composite convex optimization with backtracking , author=. Foundations of computational mathematics , volume=. 2014 , publisher=

    2014

  46. [54]

    Optimization Methods and Software , volume=

    Primal--dual accelerated gradient methods with small-dimensional relaxation oracle , author=. Optimization Methods and Software , volume=. 2021 , publisher=

    2021

  47. [55]

    Mathematical Programming , volume=

    Understanding the acceleration phenomenon via high-resolution differential equations , author=. Mathematical Programming , volume=. 2022 , publisher=

    2022

  48. [56]

    SIAM Journal on Optimization , volume=

    Generalizing the optimized gradient method for smooth convex minimization , author=. SIAM Journal on Optimization , volume=. 2018 , publisher=

    2018

  49. [57]

    Applied and Computational Harmonic Analysis , pages=

    Proximal subgradient norm minimization of ISTA and FISTA , author=. Applied and Computational Harmonic Analysis , pages=. 2025 , publisher=

    2025

  50. [58]

    SIAM Journal on Optimization , volume=

    An accelerated hybrid proximal extragradient method for convex optimization and its implications to second-order methods , author=. SIAM Journal on Optimization , volume=. 2013 , publisher=

    2013

  51. [59]

    Journal of optimization theory and applications , volume=

    Multiplier and gradient methods , author=. Journal of optimization theory and applications , volume=. 1969 , publisher=

    1969

  52. [60]

    Optimization , pages=

    A method for nonlinear constraints in minimization problems , author=. Optimization , pages=. 1969 , publisher=

    1969

  53. [61]

    Journal of Artificial Intelligence Research , volume=

    Asymptotically fair and truthful allocation of public goods , author=. Journal of Artificial Intelligence Research , volume=

  54. [62]

    Mathematical Programming , volume=

    Universal gradient methods for convex optimization problems , author=. Mathematical Programming , volume=. 2015 , publisher=

    2015

  55. [63]

    IEEE Transactions on Wireless Communications , volume=

    On integrated sensing and communication waveforms with tunable PAPR , author=. IEEE Transactions on Wireless Communications , volume=. 2023 , publisher=

    2023

  56. [64]

    arXiv preprint arXiv:2607.11878 , year=

    Optimal Parameter-Free First-Order Methods for Convex Optimization with Unknown Growth and Smoothness , author=. arXiv preprint arXiv:2607.11878 , year=

    arXiv

  57. [65]

    arXiv preprint arXiv:2603.23492 , year=

    Universal and Parameter-free Gradient Sliding for Composite Optimization , author=. arXiv preprint arXiv:2603.23492 , year=

    arXiv

  58. [66]

    SIAM Review , volume=

    The augmented lagrangian methods: Overview and recent advances , author=. SIAM Review , volume=. 2026 , publisher=

    2026

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.